Introduction

Welcome to the project page of the Deep Learning for Computer Vision Seminar. Here you can find resources that will help you develop your projects.

Projects can be developed using any of the many available deep learning frameworks. However, we encourage you to use Keras, which is easy to learn and get used to. Students can also choose whether to run their project experiments in the provided server (CPUs only), or to use your personal computers.

Slides

• Project kick off
• Oral Presentation instructions

Server Access & Setup

The instructions for the access to the server will be provided to students by e-mail.

Once you are logged in the server, you will need to setup your working environment. The simplest way will be to use a virtual environment to install your dependencies. As an example, if you are working with keras:

dlcv@imatge-dlcv:~$ virtualenv keras-env
dlcv@imatge-dlcv:~$ source keras-env/bin/activate
(keras-env) dlcv@imatge-dlcv:~$ pip install theano
(keras-env) dlcv@imatge-dlcv:~$ pip install keras

Installation in personal laptops

Here is a list of resources to install the project dependencies in your laptops:

• Docker for Deep Learning. Contains most deep learning libraries. It works for Linux, MacOSX and Windows. However, GPU support is only available for the first two.
• If you are using Windows 10 & Keras & want to have GPU support, here is a detailed installation guide.
• TensorFlow, CUDA, OpenCV Installation Guide by Teaching Assistant Andrea Ferri. Please use the issues section in his repository for any questions you may have about it.